1. Field of the Invention
The present invention relates to an image detection technology, and more particularly to an image detection technology capable of rapidly detecting the physical location of a specific object within an image through iterative algorithm.
2. Description of the Related Art
Recently, intelligent image detection technology is gaining more and more attention and can be applied to various electronic products, such as digital cameras, somatosensory interaction products and safety monitoring systems can all be equipped with the intelligent image detection technology. In these detection applications, shape information of an object plays a very important role in image detection speed and quality. More sufficient shape information, even an approximated shape, can both greatly accelerate the detection speed and improve the detection quality.
Generally, for estimation of shape information of a specific object in an image, predetermined color information can be utilized to perform pixel-by-pixel comparison, and comparison scope is gradually expanded until color values of pixels in the image are not matched with predetermined conditions. In a field of image segmentation, this manner is frequently used, especially for watershed and region growing manners. Although these manners can provide well estimation effect for objects with pure colors, these manners also have many restrictions on calculation speed, connected component and setting of threshold conditions.
However, in order to perform shape estimation for objects having complicated colors, graph-based method, such as graph-cut and normalized cut, provides an effective solution scheme in recent years. Although the foregoing methods can precisely estimate the shape of the object having complicated colors, the methods also have extremely high complexity at calculation. Similarly, model-based method, such as active contour method, also has the same problem. Therefore, the acceleration of the detection speed, detection quality improvement, restriction of connected component, insufficient flexibility and fault tolerance caused by threshold conditions that are difficult to be defined, and reduction of complexity of estimating the object having complicated colors are the main issues that the present invention is designed to overcome.